The performance of a small group (9-13) of experienced reference laboratories performing the determination of lead in blood is assessed by means of repeated analyses of lyophilized animal blood. Both interlaboratory precision and intralaboratory precision are reported for the reference laboratory group. Results indicate that at the new blood lead action level of 10 micrograms/100 ml, this group can achieve results that are accurate to within +/- 2 micrograms/100 ml. The capability of routine clinical laboratories participating in the New York State Department of Health's Proficiency Testing Program for Blood Lead from 1979 to 1991 is also examined. These data indicate a steady improvement in laboratory performance over the last 10 years. Analysis of the most recent data suggests that laboratories using anodic stripping voltammetry tend to underestimate the concentration of lead in blood at levels greater than 40 micrograms/100 ml (1.93 mumole/L). The accuracy and precision of electrothermal atomization and Delves-cup atomic absorption spectrometry for the determination of blood lead in one laboratory are compared and discussed.